On 22 September 2011 09:18, Milamber <[email protected]> wrote: > > > Le 21/09/2011 23:49, sebb a ecrit : >> On 22 September 2011 00:13, Milamber <[email protected]> wrote: >> >>> [snip] >>> netstat -tplna| grep java | grep ESTA | wc -l >>> 39 (esta conn) >>> >>> ==> 39 is the number of http requests in my test plan >>> >> I see the problem now. >> >> The httpclient cache key includes the full URL, so only reuses >> connections for identical URLs. >> >> It should match on host name, port and protocol. >> >> If you test instead with identical URLs - or loop a single URL - it >> should work the same as HC3.1 >> That's why I did not spot it earlier. >> >> I'll work on a fix now. >> > > > Your last commit fixes the problem. > Now 1 connection with second test case (bug 51863), and 100 connections > with my load test with 0% error during the test (second issue fixed too). > > Thanks.
I'm glad you spotted the issue. >>>> [I think we need to fix this before releasing 2.5.1 >>> >>> Yes sure. >>> >>> > > Any objection now to start RC1 process tonight? > No objections, go ahead. I don't think there are any other issues that cannot wait. It's sometimes tempting to add just one more fix, but then the process can slip and slip ... > Milamber > > > >>> >>>> >>>>> HTTP conversations show Keep-Alive header (HC3 and HC4) >>>>> >>>>> I suppose that is a close keep alive connection problem with JMeter. >>>>> Perhaps 2 issues are linked? >>>>> >>>>> >>>> Dunno yet. >>>> >>>> >>>> >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: [email protected] >>> For additional commands, e-mail: [email protected] >>> >>> >>> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> >> > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
